Technology

The Technology involved for Finger print Biometrics can be divided in to two major categories.

  • Image acquisition Technologies
  • Image Processing Technologies

Image acquisition Technologies :

Today we have many matured concepts of physics, which are capable of generating good quality three dimensional images. Amongst them the following are suitable to acquire a good quality image of a finger print.

  • Optical
  • Capacitive
  • Thermal
  • RF

Optical Sensors

Optical fingerprint imaging involves capturing a digital image of the finger print using visible light. In simple terms, this type of sensor is, just a specialized digital camera.

This type of sensors have limitations in imaging a unclean skin of the finger. In addition, it is relatively easy to circumvent this sensor.

Capacitive Sensors

Capacitive Sensors use the principles of Capacitance to form an image of the fingerprint patterns on the dermal layer of skin. Each sensor pixel is used to measure the capacitance at that point of the array. The capacitance varies between the ridges and valleys of the fingerprint due to the fact that the volume between the dermal layer and sensing element in valleys contains an air gap. The measured capacitance values are then used to distinguish between fingerprint ridges and valleys.

RF Sensors

RF sensors make use of the principles of medical ultrasonography to create visual images of the fingerprint. The sound waves are generated using piezoelectric transducers and reflected energy is also measured using piezoelectric materials. Since the dermal skin layer exhibits the same characteristic pattern of the fingerprint, unlike optical sensors, RF sensors can work well even if finger skin is not very clean.

Thermal Sensors

Thermal Sensors are made out of pyroelectric materials, which change their electrical property according to the temperature. A pixel array is formed using this materials, which sense the heat at all points of the finger. The heat at the rigde will be that of the body temperature, with the same at the valley will be the atmospheric temperature of the air. these temperature difference is used to differentiate between the peak and the valley of the finger print. This is measured at very high resolution as each sensor pixel is as tiny as 50 micron X 50 micron. A 3D image of the fingerprint is generated using as many as 100,000 sensing points.

Image Processing Technologies :

All fingerprint have lines which are called ridges (peaks) and valleys. In a finger, these lines bifurcate, end and form many different characteristics. Please find samples of various ridge characteristics in the image below.

These ridge characteristics like ridge island, ridge ending etc.. are called minutiae points. It has been an empirically established fact that the spread patten of these minutiae points if one fingerprint will never be the same with the spread patten of another fingerprint.

To extract the minutiae patten from the finger print image the digital image goes through a set of digital image processing techniques. The whole process can be broadly classified in to four phases namely:

  • Image Enhancement
  • Binarization
  • Thinning
  • Minutiae Detection
  • Matching and Scoring

Image Enhancement:

As it may not be possible for all the people to keep all the fingers clean all the time, so may be the image of the fingerprint. Whatever be the image acquisition technology, the contrast and completeness of the digital image generated by the sensor, still requires enhancement, to get accurate results.

Removing blurs , adding mixing pixels, improving the image contrast etc are archived by adapting Spacial methods and frequency methods for image enhancement and subjecting the image in to transformations like Fourier and Gaussian.

Binarization:

This enhanced image is a gray scale image, and all the pixels forming this image have an intensity value ranging any where between 0 to 255. Various parameters like average image pixel intensity, ridge pixel's intensity distribution, vally pixel's intensity distribution etc are analyzed to arrive at a threshold intensity value.

Then all the pixels in the image are binarized for the threshold value, to have an intensity of either 0 ( black) or 255 ( white).

Thinning:

Physically a finger print ridge ( or a valley ) will be around 300 - 500 microns.Since the image acquisition is carried out at 50 micron, the digital image of a ridge line ( or a valley line ) will be formed by around 6 - 10 pixels. The middle pixel of this ridge line maintained at intensity 255 and the rest of the pixel's intensity is changed to zero, thereby thinning the ridge line width to a single pixel line.

Minutiae Detection:

The thinned image is then analyzed to spot minutiae points. A eight pixel square is formed around the ridge pixel and this pixel square is moved along the ridge line. Any ridge pixel having more (or less) than two of the eight pixels in the square with intensity value other than 255 is noted as a minutiae point.

The location ( x,y coordinates) of all these minutiae points, ( along with angle theta, in case it is a bifurcation minutiae) stored in a encrypted file format. This is called the minutiae template file. A finger print may have around 50 - 80 minutiae points and the minutiae template will be around 800 bytes as against a 80 kilobyte image. This makes it so easy to transmit it through even thin bandwidths for comparisons.